• Robot-based 3D Scene Analysis and Understanding:

    Taking as input 3D laser scans dynamically acquired on the mobile-robot basis, we aim to analyze and understand the scenes, specifically including scene segmentation, object discovery, object recognition, function-based object classification and labeling, 3D shape abstraction and reconstruction.

  • Large-Scale Point Cloud Scanning & Processing:

    The emergence of laser/LiDAR sensors, reliable multi-view stereo techniques and more recently consumer depth cameras have brought point clouds to the forefront as a data format useful for a number of applications. Unfortunately, the point data from those channels often incur imperfection, frequently contaminated with severe outliers and noise. Our goal is to design robust consolidation algorithms for those defect-ridden point clouds from 3D scenes to significantly improve data quality.

  • LiDAR Data-Driven Scene Reconstruction:

    Rapid advances in laser scanning technology and the recent proliferation of GIS services have been driving a strong trend towards 3D reconstruction of large-scale scene models based on laser scanning point clouds. We are looking at algorithms to automatically and efficiently reconstruct large-scene scenes based on 3D LiDAR point data.

  • Digital Geometry Processing:

    Geometry processing, or mesh processing, is an area of research that uses concepts from applied mathematics, computer science and engineering to design efficient algorithms for the acquisition, reconstruction, analysis, manipulation, simulation and transmission of complex 3D models. We have been devising techniques on mesh smoothing, denoising, segmentation and reconstruction.

  • Reverse Engineering (3D Model Reconstruction):

    The reverse-engineering process involves measuring an object and then reconstructing it as a 3D model. We have been working on modeling reconstruction of industrial products, including aircraft fuselage, mechanical parts and so forth.